mass spectrometric measurement of an

The time course of both 160 and lsO product formation can be followed simultaneously, permitting rapid KIE measurements. With 180-substituted ethanol,...
1 downloads 0 Views 927KB Size
6935

J . Am. Chem. Soc. 1983, 105, 6935-6941

Permeable Membrane/Mass Spectrometric Measurement of an Enzymatic Kinetic Isotope Effect: a-Chymotrypsin-Catalyzed Transesterification K. C. Calvo, C. R. Weisenberger, L. B. Anderson, and Michael H. Klapper* Contribution from the Department of Chemistry, The Ohio State University, Columbus, Ohio 43210. Received October 18, 1982 Abstract: The 160/180 kinetic isotope effect (KIE) associated with an a-chymotrypsin-catalyzedtransestenficationwas measured by using the technique of permeable membrane/mass spectroscopy. This almost in situ method is based upon the ability of the ester product, but not water and other polar compounds, to permeate through a dimethyl silicone membrane, which separates the aqueous reaction solution from the evacuated inlet of the mass spectrometer. The time course of both l60and l*Oproduct formation can be followed simultaneously, permitting rapid KIE measurements. With 180-substitutedethanol, the KIE for the formation of both ethyl 2-furoate and ethyl 5-n-propyl-2-furoatefrom the respective p-nitrophenyl esters is initially normal and decreases with time to reach within approximately 2 min at 25 OC,pH 8.5, a constant value of 1.009 i 0.007 for the former and 0.90 f 0.02 for the latter. The large inverse steady-state KIE associated with the formation of ethyl 5-npropyl-2-furoate decreases when the temperature is raised or when the pH is lowered. The pH dependence can be fit with an apparent pK, of 7.3. The 160/180 KIE (ethyl ester oxygen) for the nonenzymatic, alkaline hydrolysis of ethyl 2-furoate was measured as 1.012 f 0.010 at 25 OC.

We undertook 160/180 kinetic isotope effect (KIE) studies to learn more about the a-chymotrypsin deacylation pathway, having found that Arrhenius plots of k,, are not linear in the temperature range of 5-40 OC for the enzyme-catalyzed hydrolysis and transesterification of p-nitrophenyl S-n-alky1-2-f~roates.~~~ These previous results are consistent with a proposal of two enzyme intermediates that, in rapid equilibrium with one another, can both turn over to release product. But the possibility that a temperature-dependent change in the rate-limiting step causes the observed nonlinearity could not be excluded. We, therefore, needed additional information concerning the number and nature of the elementary steps in the enzyme deacylation. The substitution of one isotope for another perturbs a reacting system minimally, but can alter rates when the atom involved is at the reaction locus. This rate alteration, the KIE, may provide information about transition-state structures and reaction pathw a y ~ . ~ When -~ the substitution involves an atom other than hydrogen, the effect is small, and the direct comparison of rates obtained with the different isotopic reagents is difficult, although possible.69 An alternate protocol is to mix the isotopic reagents together, thereby running the reactions simultaneously and minimizing experimental variability. The KIE can then be extracted by determining the difference in the isotopic ratios of reactants and products. This mixed isotope protocol may yield an observed KIE quite different from that obtained by direct measurement of rates in separate reactions. If the reaction is higher than first order overall, and if the isotopically separable compounds compete for the same reactant(s), then the mixed iosotopic reactants are linked, affecting the observed KIE. Thus, both protocols may yield more information than either alone.1° The mixed protocol requires a method for detecting the isotopically separable compounds-commonly, mass spectrometry. Isotopic compositions of high precision can be determined with an isotope-ratio mass spectrometer.” Because of technical requirements, only volatile substances substances can be measured. (1) Baggott, J. E.; Klapper, M. H. Biochemistry 1976, 15, 1473-1481. (2) Wang, C.-L. A,; Calvo, K. C.; Klapper, M. H. Biochemistry 1981, 20, 1401-1 408. (3) Maggiora, G. M.; Schowen, R. L. In ‘Bioorganic Chemistry”; van Tamelan, E. E., Ed.; Academic Press: New York, 1977; Vol. 1, pp 173-229. (4) Klinman, J. P. Adu. Enzymol. Relat. Areas Mol. Biol. 1978, 46, 41 5-494. (5) Melander, L.; Saunders, W. H. “Reaction Rates of Isotopic Molecules”; Wiley: New York, 1980. (6) Mitton, C. G.; Schowen, R. L. Tetrahedron Lett. 1968, 5803-5806. (7) Gorenstein, D. G. J . A m . Chem. SOC.1972, 94, 2523-2525. (8) Rosenberg, S.; Kirsch, J. F. Anal. Chem. 1979, 51, 1375-1379. (9) Rosenberg, S.; Kirsch, J. F. Anal. Chem. 1979, 51, 1379-1383. (10) Rosenberg, S.;Kirsch, J. F. Biochemistry 1981, 20, 3189-3196. (1 1) O’Leary, M. H. Methods Enzymol. 1980, 64, 83-104.

Thus, the reaction under study must involve either a component that is, or that can be, converted to a small volatile analyte. A major advantage is that large isotope enrichments are not required, easing synthetic difficulties. For example, 13C/12C KIE measurements are possible with natural abundance reagents. If a small, volatile analyte is not available, reactants and/or products and can be purified either in bulk12 or by gas ~hromatographyl~ then introduced into a mass spectrometer. In any of these techniques, the time required for sample preparation before analysis precludes the measurement of many sample points over a small reaction interval. (Mass spectral measurements without prior separation of components have been performed, however, in the case of gas-phase rea~ti0ns.l~) An entirely different approach is based upon the perturbation observed when a reaction is started at chemical but not isotopic equilibrium.Is If a fluorescing or light-absorbing compound is involved in the reaction, then the perturbation that follows the addition of isotopically substituted reagents can be measured spectrally, and the KIE can be determined with no need for a mass spectrometer. This method is limited, however, to those reactions in which there is a measurable equilibrium. While studying a-chymotrypsin catalysis with the mixed isotope KIE (ethanol oxygen) in protocol, we had looked at the 160/180 the transesterification of p-nitrophenyl 5-n-propyl-2-furoate (eq 1, R = CH3CH2CH,). There was a large, >lo%, inverse KIE R//\iC//O 0 N -‘02 ‘

+

CH,CH,OH

(I8k > 16k)associated with the steady-state formation of the ethyl ester product.I3 Surprisingly, the initial product isotopic ratios (normalized to that of the starting ethanol) were normal, suggesting an initial normal KIE (I6k > l*k) that decreased with time to become inverse. The novelty of these results prompted us to attempt improvement in the precision of the data and to obtain better delineation of the time dependence for the apparent KIE. To achieve these goals we have developed a new method based upon permeable membrane/mass spectroscopy (PM/MS). A description of the method together with its application to the (12) Sawyer, C. B.; Kirsch, J. F. J . A m . Chem. SOC.1973, 95, 7375-7381. (13) Wang, C.-L. A,; Trout, C. M.; Calvo, K. C.; Klapper, M. H.; Wong, L. K. J . A m . Chem. Soc. 1980, 102, 1221-1223. (14) Kwart, H. Acc. Chem. Res. 1982, 15, 401-408. (15) Cleland, W. W. Methods Enzymol. 1980, 64, 104-125.

0002-7863/83/1505-6935$01.50/0 0 1983 American Chemical Society

6936 J . Am. Chem. SOC.,Vol. 105, No. 23, 1983

Caluo et al. of the response curve and is proportional to V,,,,,.

2700k

b~ = ~ ~ R o H ( E o ) ( C H ~ C H Z O H )

2100

where a,the proportionality constant, can be obtained by calibrating the instrumental response to known concentrations of the ester. It can be shown (see below) that when I 6 0 and I8O reactions occur together, the ratio of steady-state slopes is

!

2 1800

I6b1 -=I8bl

J.J

"! I

.

1

a 900

300

(3)

.! 2

1

I

I

3

4

Time ( m i n )

Figure 1. a-Chymotrypsin-catalyzed transesterification of p-nitrophenyl 2-(5-n-propyl)furoate to the ethyl ester: ( 0 )mass spectral peak intensity of the I6O ethyl ester parent ion; (B) mass spectral peak intensity of the "0 ethyl ester: Conditions are described in Table 11; R for this experiment was 0.94. The lines drawn through the points are arbitrary and are used only to aid in visualization.

measurement of reaction 1 has been published earlier.16 We report here the PM/MS measurement of the KIE's for the a-chymotrypsin-catalyzed transesterification of p-nitrophenyl 2-furoate and p-nitrophenyl 5-n-propyl-2-furoate and the temperature and pH dependence of the KIE for the latter reaction. The time dependence of the apparent KIE that we reported previously is verified, as is the large steady-state value. Methods and Materials We have previously described the methodology for the measurement of reaction rates by PM/MS,16 so that only a brief description is given here. A dimethyl silicone polymer membrane of nominal thickness 20 pm (General Electric, Schenectady, N Y ) was clamped between two stainless steel blocks, forming a reaction cell with the membrane as its bottom (area = 1 d ) . This membrane excludes water almost totally, while the ethyl ester product can permeate across. The temperature was maintained by water pumped through the upper block, and the cell was connected through the lower block and via stainless steel tubing to the source of the MS, a Du Pont Model 21-490 single focusing, magnetic sector machine. The MS was operated in a selected ion mode by varying the accelerating voltage to scan two peaks, the parent (or base) and the corresponding 180-containingM 2 peak, for approximately 3 s each. The following procedure was typical for measuring reaction 1. Buffer (0.42 mL of 0.1 M K2HP0,), ethanol (13 p L of a mixture containing approximately 1/1 ['60]-/['80]ethanol), and a-chymotrypsin (50 pL of enzyme in 0.001 M HCI; Worthington, Freehold, N J ) were allowed to come to temperature in the reaction cell; the reaction was initiated by addition of the p-nitrophenyl ester substrate in 20 pL of dimethylformamide. After an initial lag, the two signals corresponding to the l6O and I8Oethyl ester products rose linearly with time, e.g., Figure 1. We have shown that the time response for detection of the ethyl ester product is fit by the equation

+

I is the signal intensity; al, a,, and a3 are constants from the time dependence of ethyl ester permeation across the membrane and can be determined in an independent experiment. The constant b2 corresponds to the first-order rate constant for approach of the enzymatic reaction to steady state. A reliable value for this constant cannot be obtained by fitting data to eq 2, since under our present experimental conditions a3 (a measure of the ethyl ester permeation rate across the membrane) is much smaller than b2.I6 The constant bl is the slope of the linear portion

(16) Calvo, K. C.; Weisenberger, C. R.; Anderson, L. B.; Klapper, M. H. Anal. Chem. 1981, 53, 981-985.

I6kRoH '6a(CH3CH2'60H)

I6kRoH

=R " ~ R o H ~ * ~ ( C H ~ C H ~ ' ~I8kRoH OH)

(4)

If the value of R is known, then the apparent steady-state KIE ( ' 6 k ~ o ~ / ' 8 k ~can o H )be calculated. All constants reported were obtained by least-squares fitting. The standard deviations reported refer to the fit of experimental points to the theoretical curve and are not meant as estimates of absolute error. Since ethanol (0.5 M) is in large excess over substrate (0.8 mM) and enzyme (0.08 mM), the 160/'80 isotope ratio of the ethanol remains essentially constant during the reaction. Rather than measure this ratio, we chose to determine R directly, for a number of reasons: the difficulty in working with the ethanol mass spectrum; the mass discrimination that arises from scanning by varying the acceleration voltage;" possible isotope effects associated with fragmentation in the MS; a possible isotope effect in the permeation of ester across the membrane; the simplicity of the procedure we devised for determining R. The same ethanol mixture used in a transesterification reaction was quantitatively converted to the ester-either ethyl 2-furoate or ethyl 5-n-propyl-2-furoate-by the reaction with the acid chloride. Ethanol, 3 pL, was mixed with 300 p L of dry pyridine and with 50 p L of the appropriate acid chloride. The mixture was held at 40 "C for 10 min, when the reaction was complete. Gas-liquid chromatography through a 6 ft by in. 0.d. stainless steel column packed with 3% OV-255 (Altech) on Chromosorb Q,60-80 mesh (Johns-Manville), completely separates ethanol from all other components in the reaction mixture. No ethanol nor any other product other than the furoyl ethyl ester could be detected, indicating >99% reaction. An aliquot of this mixture was added to the MS cell, which contained the buffer solution for the transesterification reaction. The plateau intensities of either the parent or base peaks for both isotopic esters are the required measure of R. The procedure just described is a one-step calibration of both the P M / M S system and the ethanol isotope ratio and was followed before each series of kinetic runs on any particular day. In support of the procedure's validity, we have shown that the measurement of R by G C / M S on the same mass spectrometer gave a value identical within experimental uncertainty with that determined by PM/MS. The ethanol/acid chloride reaction mixture was applied directly to the GC column, and the effluent was monitored in the single-ion mode for the two parent peaks ( m / e = 140 and 142 for ethyl 2-furoate). The areas under the two chromatographic peaks were integrated numerically, and their ratio was taken as the measure of R. With ethyl 2-furoate, and one particular ethanol mixture, we obtained 0.855 f 0.008 (GC/MS) and 0.861 h 0.003 (PM/MS). The acid chlorides of 2-furoic and 5-n-propyl-2-furoic acid were made by reaction with thionyl chloride. For the furoyl chloride, 200 mmol of the acid was dissolved in 20 mL of SOCll and refluxed for 1 h. The excess SOCI, was boiled off, and the acid chloride was distilled at 172 "C (lit. bp 173-174 "C). The product was a clear liquid with a mass spectrum consistent with the proposed structure. The 5-n-propyl-2-furoyl chloride was synthesized in a similar manner. Both acid chlorides were stored at 4 "C in a desiccator. 5-n-Propyl-2-furoic acid was synthesized from n-propylfuran as described previously.' The ethyl esters of both acids were prepared by reaction of the acid chlorides with ethanol. The ethyl ester of furoic acid, which was recrystallized from water/methanol (70/30 v/v) at -78 "C, had a mp of 32-33 OC (lit. mp 34 "C). The ethyl ester of 5-n-propyl-2-furoic acid was isolated as an oil and was deemed pure by G C / M S . The p-nitrophenyl esters of the two acids were synthesized as described previously.' ['80]Ethanol was synthesized by the method of Sawyer.18 All other reagents were purchased from commercial sources and used with no further purification, except p-nitrophenol and 2-furoic acid, both recrystallized from ethanol.

Results Kinetic measurement by PM/MS is based on the permeation of at least one component in a reacting system through a membrane barrier into the inlet of a mass spectrometer. When enzymatic reactions are studied, the advantage of the dimethyl (17) Coggeshall, N. D.; J . Chem. Phys. 1944, 12, 19-23 (18) Sawyer, C. B. J . Org. Chem. 1972, 37, 4225-4226.

J . Am. Chem. SOC.,Vol. 105, No. 23, 1983 6937

PM/MS Measurement of Chymotrypsin KIE silicone polymer membrane is its impermeability to water, resulting in an efficient one-stage purification for sufficiently apolar analytes. Since the magnitude of the permeant component's mass spectrum is directly proportional to its concentration in solution, time-dependent changes in that concentration can be followed, when solute movement across the membrane is also taken into account.I6 We consider a nonreacting system first. When ethyl 5-npropyl-2-furoate is added to a buffer (0.1 M phosphate, pH 8.5) solution in the PM/MS cell, there is a rapid increase in the magnitudes of the parent ( m / e = 182) and base ( m / e = 154; loss of ethylene) ion peaks. The half-life for this increase, approximately 25 s at 25 OC under our experimental conditions, is related to 12/D, the ratio of the membrane thickness squared and the diffusion coefficient of solute in the membrane. Subsequently, there is a slow first-order decline of the signal with tlj2 in the range of 3-12 min between 18 and 38 "C ( t l 1 2= 8.2 min at 25 "C). This decay, due to permeation loss of the ester, is so much slower than the initial buildup that it can be assigned a first-order rate constant to represent the permeation process. When a mixture of 160/'s0 (ethyl) esters is added to the cell, the permeation loss of both can be monitored at either m / e = 182 and 184 or m / e = 154 and 156. Since first-order constants are concentration independent, the isotopic ratio of the starting mixture need not be known, and the permeation isotope effect is the ratio of first-order constants for I6O and I8O compounds. For ethyl 5n-propyl-2-furoate, the isotope effect was 1.006 f 0.006 over the temperature range 18-38 OC. A small or experimentally undetected effect would be expected. We consider next the nonenzymatic saponification of ethyl 2-furoate, a first-order reaction. With the ester in dilute alkali, the first-order signal decay for parent ( m / e = 140) and base ( m / e = 112) ions results from two independent processes, hydrolysis and ester loss through the membrane. Thus, the hydrolysis rate constant can be obtained from the difference in decay times for alkaline and neutral (nonreacting) conditions. Saponification of the ethyl 2-[160,180]furoatemixture yields a hydrolysis rate constant for each isotopic species. Since the two reactions are first order, the isotopic ratio in the starting mixture need not be known, and the KIE is the ratio of these two constants. The observed KIE for ethyl 2-furoate is marginal at 1.012 k 0.010 (pH 10.0, 25 "C, 16k = 0.1244 f 0.003 m i d ) , consistent with reported results: 1.0091 0.0004 for saponification of methyl formateI2 and 1.0062 f 0.0006 for methyl b e n ~ 0 a t e . l ~ Finally, in the a-chymotrypsin-catalyzed transesterification of p-nitrophenyl furoyl esters, the ethyl ester is the product (eq l), and the reaction is followed as an increase in signal intensity. (We have not been able to detect permeation of the p-nitrophenyl substrate nor of the acid formed in the competitive hydrolysis of the acyl enzyme. Under our experimental conditions, approximately 25% of the furoyl and 50% of the n-propylfuroyl reaction is diverted to hydrolysis.) Ethanol does pass through the membrane, although less readily than the ester; due to its low mass it does not interfere with detection of the ester. The results of one experiment with a 160/180 mixture of ethanol is shown in Figure 1. The initial lag in both curves is due primarily to the transport of solute through the membrane.16 The slopes of the linear portions of the curve are proportional to the enzymatic k,,, and their ratio is proportional to the steady-state '60/180 KIE (eq 4) as we show in the Appendix. Note from the results in Figure 1 that steady state is achieved in less time (